The Great Aksumite Decorated Stelae: Architectural Characteristics, Functions, and Meanings

The article presents a review of the architectural and iconographic features of the big decorated Aksumite stelae in the May Ḥǝǧǧa stelae field at Aksum. Their location in the urban setting of ancient Aksum is scrutinized alongside their ceremonial and ideological function. The origin and meaning of the different features and decorative patterns characterizing the stelae are focused upon. It is suggested that these attributes may reflect ideological traits regarded as crucial by the Askumite kings and the elite in the first centuries CE. Moreover, it is demonstrated how some of these features are rooted in the local traditions, while others are related to the intense interactions the ancient Aksum had with neighbouring regions, such as the Mediterranean area and South Arabia. In particular, a new interpretation is proposed for the very distinctive outline of the top of the monuments: it may have been shaped after a specific type of shield also occurring in Meroitic and Post-Meroitic Nubia. Finally, it is suggested that the role these monuments may have played in ceremonies aimed at shaping the Aksumite identity.


Introduction
In studying Aksumite culture, the largest decorated stelae of the May Ḥǝǧǧa stelae field immediately appear to be of clear, central significance. The huge monuments, up to 30 m high, weighing up to 520 tons, are presumed to be from the late third-mid fourth century CE. 1 Located in the centre of modern Aksum, facing one of the town's main squares, they are not only some of the most visible monuments of obvious tourist interest, but still play an active part in the lives of modern Aksumites (Fig. 1). They provide the setting for rituals such as the dämära, part of the Mäsqäl celebration for the Ethiopian New Year on 16 September, the Ṭǝm ät or the festival of St Mary of Zion. 2 The earliest detailed and systematic description of the large decorated stelae is a major scholarly contribution made by the 1906 Deutsche Aksum-Expedition (DAE) led by Enno Littmann. More recent investigations have increased the knowledge of the largest stelae of Aksum considerably, but Daniel Krencker's description remains the reference study for these fascinating monuments. 3 Their funerary meaning had been suggested at the very outset, but only confirmed after the investigations conducted in the May Ḥǝǧǧa stelae field by the British expedition led by Neville Chittick in 1973 In recent years, the Italian-American investigations conducted at the site of ʿOna Ǝnda Abboy Zäg w e, on top of the Betä Giyorgis hill, north of Aksum, under the direction of Kathryn A. Bard and Rodolfo Fattovich, have shown that the occurrence of funerary stelae in Aksum goes back at least to the third, perhaps late fourth century BCE, at the beginning of the Proto-Aksumite period. 5 Moreover, it has been suggested that the Aksumite funerary stelae may be rooted in a very ancient tradition of North-East Africa going back to prehistoric times, perhaps to the fifth to fourth millennium BCE. 6 In fact, the stelae of the third to second millennium BCE cemetery discovered at Mahal Telignos in the Sudanese-Eritrean lowlands are also expressions of this tradition and the geographically closest to the stelae of the Ethiopian-Eritrean highlands. 7 In recent years, a chronological and spatial link between the stelae of the lowlands and those of the Proto-Aksumite was discovered in the pre-Aksumite cemetery of Sǝglamen, not far from Aksum, dating to the first millennium BCE. 8 Nevertheless, the Sǝglamen finds are certainly not unexpected, taking into account the badly broken monolithic stelae near the funerary pits in the Daʿro Mikaʾel pre-Aksumite cemetery investigated by Francis Anfray at Yǝḥa in 1960. Unfortunately, thus far the Yǝḥa stelae have been almost completely overlooked, due to their registering being only a very slim mention in a publication footnote. 9 Following the Proto-Aksumite stelae, others dating back to fully Aksumite times were erected at Betä Giyorgis: featuring among them are the only double stela discovered to this date. 10 Other stelae were erected in downtown Aksum in the first centuries CE, both undressed and dressed, round topped, pointed or featuring elaborate, sculptured decoration. 11 During the British archaeological expedition directed by David W. Phillipson, a systematic study was also conducted on the quarries of the stelae, located on the southern and eastern slopes of the Gobǝdra hill, west of Aksum, and on the techniques used for their preparation, transportation, and erection. 12 Further studies have been devoted to these topics more recently. 13

Stelae, Platforms, and Ceremonies
In the Aksumite cemeteries the stelae were not just intended to mark the presence of tombs. Similar to the modern period, they were actually used as a set, a location where activities took place. They were erected on top of rubble platforms delimited by walls and accessed through stairs, as shown by excavations conducted at Aksum both in the stelae fields of May Ḥǝǧǧa and Dungur, 14 and at ʿOna Ǝnda Abboy Zäg w e, atop the Betä Giyorgis hill. 15 It is contended here that the platforms should not be considered merely as stabilizers for the huge upright stones: when the platforms were progressively extended, 16 the carefully paved and repaved tops became living floors, where activities took place. Of course, the reconstruction of such activities may be difficult at May Ḥǝǧǧa, the main stelae field of the capital city, as the area's continuous frequentation to the present day has destroyed most traces on top of the platforms. Despite problems caused by agricultural activities in the area, at ʿOna Ǝnda Abboy Zäg w e, Italian-American explorations showed that some concentrations of pottery, usually large basins, cups, and beakers, could be identified on the top of the Proto-Aksumite platforms. 17 These materials may provide some insights into the funeral ceremonies that took place there. The number and type of vessels discovered in some assemblages suggest that not only were ritual offerings performed, but drinks and perhaps food may also have been partaken of close to the stelae and the tombs. Interestingly, a täzkar, a kind of funerary meal, still takes place in traditional Ethiopia during a funeral and on the fortieth day afterwards. This ceremony may have ancient Aksumite roots for the term täzkar may be that which is referred to in a Gǝʿǝz inscription on a schist slab from Mäṭära published by Lanfranco Ricci. 18 Moreover, the materials found on top of the platforms at ʿOna Ǝnda Abboy Zäg w e also show the prolonged frequentation and continuous use of these structures, sug-14 About May Ḥǝǧǧa see Chittick 1974, 164-167, 175, 180-182, figs. 4-5, 12-13, pl. VII, IX b;Fattovich et al. 2000, 51;Leclant 1959, 4-5, 10, pl. V-VII;Munro-Hay 1989, 55, 83-90, 152-155, 329-330, figs. 6.4, 6.16-18, 6.25, pl. 6.11-12, 6.23;Phillipson 2000, II, 480, fig. 417. About Dungur see Fattovich et al. 2000, 51, n. 37, see also Ayele Tarekegn and Phillipson 2000, 227, fig. 206 a. 15 Fattovich andBard 1991, 48-55, 68-69, pl. I, fig. 1;Fattovich and Bard 1993, 7, 10-13, 19-26, figs. 2, 11;Fattovich and Bard 1995, 58-61, fig. 5;Fattovich and Bard 1997, 65, 68;Fattovich and Bard 2001, 5-12, 15;Fattovich and Bard 2002, 9-12;Fattovich et al. 2000, 50-51. 16 Chittick 1974, 164-166;Munro-Hay 1989, 152-155. 17 Fattovich and Bard 1993, 24-26;Manzo 2014, 10. 18 Ricci 1991Ricci , 1298Ricci , 1306Ricci , 1310. On this see also Bausi 2013, 180-181. gesting perhaps the continuity of the funerary cult over prolonged periods of time. 19 The fact that libations and offerings took place in front of the stelae is also confirmed by special fixed installations, such as the stone offering basins still visible, for example, close to some Aksumite stelae at Yǝḥa, 20 and carved on the base slabs of some of the big sculptured stelae in the May Ḥǝǧǧa cemetery, where in some cases the shape of a fluted two-handled vessel is reproduced (Fig. 2). 21 Perhaps, the term 'base slab' itself, used in recent publications, is insufficient in describing the ritual function of these slabs, and would be better served by the term 'Altarplatte' used by Krencker. It has been noted that the shape of the vessels sometimes carved on the 'Altarplatten' is reminiscent of the kylix or of the krater, 22 the former referring to a cup used for drinking wine and the latter a bowl used for serving wine in the Mediterranean regions. This may well have been the case in Aksum also, where wine was imported and perhaps produced locally at the time when the stelae with sculptured decorations were erected. 23 Perhaps the frieze featuring the vine-leaf with grape bunches decorating the edges of the 'Altarplatte' of Stela 3 is making reference to precisely that. 24 Unfortunately, the reuse of the original stone offering basins carved on the 'Altarplatte' of Stela 3 as mortars has completely effaced their original appearance (Fig. 3). 25 It is apparent that ceremonies were performed in front of the stelae. As the big stelae with sculptured decoration were probably used to mark the location of royal graves, these were official ceremonies related to the ideology of the Aksumite State. The dimensions of the stairs leading to the top of the platform at May Ḥǝǧǧa, wider than 3 m, 26 suggest that a certain number of people may have attended and/or performed on the top of the platform, while others, perhaps the majority of those attending the ceremonies, remained at the large area in front of the platform, below the area where the 19 See e.g. Fattovich and Bard 1993, 26. 20 Krencker 1913, 2, 79, fig. 2, 163; see also Fattovich 1987, 47;Manzo 2014, 10. 21 Krencker 1913, 13, 16, fig. 25, pl. III-IV;Phillipson 1997, fig. 28;Manzo 1999, 352-354, figs. 4, 5;Manzo 2014, 9-10. 22 For the former see Chittick 1974, 163, pl. VI a. For the latter see Manzo 1999, 354. 23 Manzo 1999, 355-356. 24 Krencker 1913, 21-22, pl. VI;Phillipson 1997, 32, fig. 35;Manzo 1999, 348, 354-355, fig. 1 a-b. 25 Phillipson 1997, 32, fig. 36. 26 Leclant 1959, 10;Munro-Hay 1989, 55, 59, figs. 6.1-4; see also Contenson 1959, 27-28, 34;, II, 480.

The Great Aksumite Decorated Stelae
Aethiopica 23 (2020) 11 rites took place. 27 A few test pits in the area in front of the big sculptured stelae excavated by Ugo Monneret de Villard in the mid-1930s, 28 and later excavations seem to confirm that no ancient structures occur there, 29 suggesting that the space may have been empty originally, forming perhaps a kind of plaza or square, in ancient times as it is today (Fig. 4). Of course, people of different ranks may have had different kinds of access to the ceremonies: few persons performed on and near the 'Altarplatte', while others watched the ritual from top of the platform and from the area in front of it. If it is accepted, as has been recently suggested, 30 that the 'Altarplatten' of the three largest stelae of the May Ḥǝǧǧa stelae field were at a level higher than the living floor on top of the platform, on a kind of podium, then it is possible to distinguish archaeologically at least three different grades of access to the ceremonies: from the area in front of the platform, from the top of the platform, and from the podiums. At any rate, most of those attending the ceremonies gathered at the space in front of the platform. Thus, the majority of the people had at most a visual access to the rite, serving perhaps to reinforce the social hierarchy by emphasizing the sacredness of the rite itself. Interestingly, the orientation of the large, decorated stelae and the location of their 'Altarplatten', when still in place, confirm that they were conceived to be observed from south, that is, from the area in front of the platform. 31

Stelae and Topography at Aksum
It should be noted that the empty area in front of the platform, where the big decorated stelae were erected, may also have been central to the topography of the ancient capital city, for two main roads, possibly roughly contemporary with the big decorated stelae, may have converged there (Fig. 5): the first road from the south, following the foot of the May Qoḥo hill, the second from the south-west. The presence of the second road is suggested by the orientation of some of the monumental structures, perhaps palaces, recorded by the DAE, 32 that were perhaps linked to the May Ḥǝǧǧa stelae field by a processional road. 33 Indeed, the road lining the foot of the May 27 Phillipson 2000, I, 161. 28 Monneret de Villard 1938, 7-8. 29 Leclant 1959, 5-6. 30 Poissonnier 2012, 70-72. 31 Phillipson 1994, 193;2000, II, 478. 32 Krencker 1913, 107-124. 33 Fattovich et al. 2000, 43;fig. 11 c. Qoḥo was a processional way too, monumentalized by a row of stone thrones with royal inscriptions erected along it. Moreover, the map published by the DAE and the photos taken by the same show that the inscription of King ʿEzana-located now in the small archaeological park where some of the objects recovered during the road constructions, carried out at the time of the Italian occupation, were collected-may have been originally erected near this way. 34 A path existed there when the DAE visited the site, and, according to the tradition, was still used as a processional way during the ceremonies of coronation in the fifteenth-sixteenth centuries. 35 To follow the processional way, as was perhaps also the direction in ancient times: it leads precisely to the square in front of the platform and of the big decorated stelae. Unfortunately, not enough data are available for a full reconstruction of the ceremonies that took place there and what ideas may have been shared with the attendees. Nevertheless, as the monumental location of these ceremonies was created and sponsored by royal authority, it may be safely established that these ideas and concepts were crucial to royal and state ideology. In such a context, it is very likely that all the components of the location itself, including the shape and decorations of its pivot element, namely the stelae, had a crucial function in delivering an intended message to the people. As a result, it is quite likely that all these components, including the shape and decorations of the stelae, were carefully chosen.

The Decoration of the Stelae and its Meaning
As is well known, the sculptured decoration of the big stelae represents multi-storeyed buildings featuring a door at the base, the 'monkey's heads' typical of Aksumite architecture, and several types of windows arranged in superimposed rows (Figs. 6, 7). 36 It is forwarded here, that this is not only related to the widespread symbolism of the grave as a house, 37 which occurs in many human cultures. The Aksumite decorated stelae do not represent a mere common house. Marking royal graves, they represent monumental royal palaces, and the palace is a symbol of the authority of the king himself. 38 This is quite evident on the other side of the Red Sea, in the south of 34 Krencker 1913, 47-48, fig. 90;Monneret de Villard 1938, 16-17;Phillipson 1997, figs. 195, 208, 209, 212; see also Fattovich et al. 2000, fig. 11 c. 35 Littmann 1913Monneret de Villard 1938, 16-17, 54-56, 63-64. 36 Krencker 1913, 7-10;Phillipson 1994, 197-200, 208. 37 See Chittick 1974, 201. 38 Manzo 1995, 165. Aethiopica 23 (2020)  13 the Arabian Peninsula, where the royal palaces bore proper names often used in the inscriptions instead of the name of the capital city and also occurred in the royal title. 39 The suggestion here is that the palace had a similar symbolic importance also in the Aksumite royal ideology. Incidentally, two inscriptions from southern Arabia, that refer to an Aksumite military involvement in that region in the second half of the third century CE, mention the proper name of a royal palace of the Aksumite king Gdrt, Zrrn, and in the texts a parallelism is established with the Sabean royal palace, Slḥn, at Mārib. 40 At Aksum, the stelae reproducing palaces are associated with large tombs (see for instance the Mausoleum) which may have been intended as multiple tombs for several members of the princely and royal families. 41 Thus, the palace may have been regarded as a symbol of the dynasty and not simply that of a single ruler. This appears also to concur with what is known of other elite tombs, apparently always intended to be family or lineage graves, such as Tomb 2 at ʿOna Ǝnda Abboy Zäg w e on top of the Betä Giyorgis hill, which was progressively enlarged to accommodate more burials. 42 A similar, progressive enlargement may be perceived in the so-called Tomb of Bazen, at the foot of the May Qoḥo hill, where, as in the case of Tomb 2 at ʿOna Ǝnda Abboy Zäg w e, the original access through a vertical shaft was replaced by a stepped approach to which additional loculi were possibly added. 43 The relation between stelae and family or lineage may be expressed by the inscription on the well-known stela of Mäṭära, dedicated by ʾGZ 'to his father', 44 possibly an ancestor. This suggests that the stela of Mäṭära, if funerary, may have been related to a family or lineage tomb, or to more than one grave. It should be stressed here that the funerary function was recently questioned regarding the stela of Ḥǝnzät, bearing a text somehow parallel to the one of the stela of Mäṭära, 45 but this will remain a matter for debate until a proper archaeological investigation can identify the context where these stelae were originally erected. At any rate, at least at Aksum, in the third and fourth centuries CE, the stelae, a funerary monument rooted in a very ancient cultural tradition of North-East Africa and there often asso-39 Ibid. 40 Robin 1989, 153;Academia Inscriptionum et Litterarum Humanorium 1920, 323-334, no. 308, l. 14, no. 308 bis, l. 13; see also Marrassini 2014, 29-32. 41 Phillipson 2000, II, 477. 42 Fattovich andBard 1993, 13;, 56. 43 Phillipson 19972012, 149. 44 Bernand et al. 1991, 318-319, no. 223;Kropp 2006, 326;Ullendorff 1951. 45 Yohannes Gebre Selassie 2014 ciated with multiple, possibly family or lineage, tombs, were dressed to reproduce a physical manifestation of the royal power and of the dynasty, that is, the palace. Taking into consideration their crucial ideological meaning, the fact that at least one of the large decorated stelae may have been intentionally destabilized and ultimately toppled in Aksumite times, as has been suggested recently, 46 may well be related to political, religious and/or perhaps dynastic troubles, an hypothesis to be seriously considered, but which remains nevertheless tantalizing for historians until more textual and archaeological evidence is available.
If the decorated stelae can be confidently considered as a highly symbolic reproduction of monumental buildings, a single detail of the shape and decoration of their tops still needs further explanation. The apex of the Aksumite stelae with architectural decorations are actually characterized by recessed sectors which feature cuttings for the attachment of some kind of metal plaques or standards (Figs. 6, 7). 47 These were fixed by means of iron pegs still partially visible in their original positions. 48 Given their position on top of the decorated stelae marking the royal tombs, the symbols represented on these plaques were surely highly meaningful for royal Aksumite ideology. Krencker suggested the represented symbol was the disk with the crescent, also characterizing pre-Christian Aksumite coins, 49 and carved on the top of the above-mentioned stela of Mäṭära. 50 On the basis of the distribution of the metal pegs on the recessed surface where the plaques were fixed, Gus W. Van Beek hypothesized that, at least in some cases, the cross was represented at the apex of the decorated stelae, in the meantime suggesting that some of the largest stelae may date to the period following the adoption of Christianity by the Aksumite kings. 51 Other decorations were carved on top of the back of Stela 3 and of Stela 4: they consist of a round, disk-shaped element. 52 As two spears are represented on the front of Stela 4 (Fig. 7), 53 Krencker's suggestion that these rounded elements may be shields could indeed be valid. 54 46 Poissonnier 2012, 81-83. 47 Phillipson 2000, I, 136, fig. 112 a. 48 Krencker 1913, 21, fig. 38. 49 Krencker 1913, 24, fig. 44. 50 Bernand et al. 1991, 318, no. 223;Kropp 2006, 324, fig. 2. 51 Van Beek 1967, 118, pl. 3. 52 Chittick 1974Krencker 1913, 21, fig. 41;Phillipson 1997, fig. 37. 53 Chittick 1974, 162, fig. 3;Phillipson 1997, 184, figs. 264, 265; see also Krencker 1913, 27-28, fig. 53; Monneret de Villard 1938, 37-38. 54 Krencker 1913, 21. See also Conti Rossini 1928Phillipson 1997, 32, n. 61. More recently, a rounded cast brass plaque with the frontal representation of a human head surrounded by a Gǝʿǝz inscription was found in the Tomb of the Brick Arches (Fig. 8), an elite family tomb close to the big decorated stelae in the May Ḥǝǧǧa stelae field. 55 Its measurements are compatible with those of the recessed sectors on the apex of some stelae. Therefore, Phillipson proposed this may have been one of the plaques adorning the apex of the decorated stelae. Agreeing with the latter it is also possible that the round element under the human face represented on the plaque is not a torque, as pretended by Phillipson, but the 'knotted' tails of two snakes with the curved elements in the upper part of the figure also being snake bodies: thus, the human face should be interpreted as the head of Gorgon. 56 Indeed, such a reference to the Mediterranean iconography is not surprising in the royal Aksumite milieu, where Greek was used as an official language, 57 and Hellenistic vine-leaf friezes decorated objects and monuments related to the Aksumite elite such as 'Altarplatte' referred to above of Stela 3. 58 In the Aksumite context, the adoption of the iconography of the frontal head of Gorgon may have been favoured by the ideological meaning traditionally attributed to snakes in northern Ethiopia. 59 Nevertheless, the fact that the frontal head of Gorgon was an imperial symbol for the Mediterranean context may well have been deemed appropriate at Aksum: particularly in the Mediterranean this kind of representation widely occurs as a central element in the decoration of the shields. 60 As highlighted above, shields may have been carved on top of two Aksumite stelae, therefore it is reasonable to consider the top fronts of the stelae with their recessed sectors and metal plaques to be somehow related to representations of shields.
It should be emphasized that the shape of the large decorated stelae, whose round tops are demarcated by single or double concavities on their narrow sides (Figs. 6, 7), 61 remains unexplained. In the light of what has just been forwarded, it could be appropriate to compare this very distinctive shape with that of leather shields of the Nile valley. A leather shield characterized by its round upper part demarcated by concavities on the narrow sides near its base was discovered in a Meroitic assemblage at Aniba (Fig. 55 , I, 97-100, figs. 79, 80. 56 Manzo 2012, 430-431, fig. 2. 57 Conti Rossini 1928Marrassini 2014, 29. 58 Manzo 2012 Manzo 2014, 13-15, 17-21, see also Appendix A to this article. 60 Manzo 2012, 431. 61 Krencker 1913, 11, 13, 16, 20-21, figs. 22, 26, 33, 37-38. 9a) 62 and perhaps a similar, but, unfortunately, rather badly damaged object was collected in a Meroitic context at Sedeinga. 63 In later times in the Middle Nile valley this type of shield was even more popular. Shields of this type were collected in Post-Meroitic assemblages, in most cases later than the Aksumite decorated stelae, dating from the fifth-sixth centuries CE, in the Fourth Cataract region (Fig. 9b) 64 and in Lower Nubia, at Qaṣr Ibrīm (Fig. 9c), 65 at Qustul, 66 and at Ǧebel Adda. 67 Moreover, these shields were represented on late first-millennium CE Early Christian ceramics from Faras, where they are seen to be worn by a man, perhaps bearing a spear (Fig. 10), 68 from Karanog, 69 and from other Lower Nubian sites. 70 These comparisons clearly show that the shape of some shields from Nubia is similar to the outline of the apex of the Aksumite stelae, therefore, it can also be proposed that the typical outline of the decorated Aksumite stelae represents a shield. At the moment it is not possible to determine if this type of shield was first adopted in Nubia and later on in Ethiopia, or vice versa. The first adoption could be the one in the Aksumite context, as most of the Nubian examples appear later than the Aksumite stelae. The absence of the leather shields being represented on top of the decorated stelae from contemporary Aksumite archaeological assemblages may well be explained by the environmental conditions of the Ethiopian-Eritrean highlands, which are less favourable to the preservation of organic materials than the extremely dry ones on the desert fringes of the Nubian Nile valley.
The fact that the outline of the top of the largest decorated stelae may represent the shape of a specific type of shield could also suit the framework of the Aksumite royal ideology, in which military elements were surely embedded and crucial. The importance of the military aspect of the Aksumite royal ideology is demonstrated not only by the above referred apex of Stelae 3 and 4 with the rounded shields (associated in the case of Stela 4 with two spears), but also by the well-known correspondence between Mäḥrǝm, the Aksumite god patron of the king, and Ares, the Greek god of war, in Aksumite royal inscriptions, where the Aksumite king also claims to be his son. 71 Indeed, the fact that Mäḥrǝm/Ares was characterized by the Aksumites as a warrior is made evident by the epithet 'unconquered' conferred on him in the inscriptions in Greek and Gǝʿǝz. 72 The military connotations of the Aksumite kings may be also expressed on coins, where the king is often represented with what may be a spear in his hands. 73 This seems to be the case in a type of gold coin of King Mḥdys, a fifth-century CE ruler, bearing the representation of the king, standing with a long spear and a rounded shield. 74 Moreover, in Chamber B and Loculus H of the Tomb of the Brick Arches, a great number of iron spear heads was discovered, 75 confirming that the military connotation and related attributes were felt as crucial by the Aksumite elite in the first centuries CE. These archaeological and iconographic elements have also led to suggestion that shields and spears could be considered as emblems of the Aksumite kingly office. 76 As attempted above, the type of shield having the same shape as the apex of the big Aksumite decorated stelae has been documented in several assemblages of the Nubian Nile valley, dating from Meroitic to Early Christian times. This is not the only element among the weapons recorded in the May Ḥǝǧǧa stelae field related to finds from the Nubian-Sudanese Nile valley. Chittick collected a fragmentary stone thumb ring of an archer, similar to those widely occurring in Meroitic and Post-Meroitic assemblages in the Nubian-Sudanese Nile valley, from a disturbed site near the Tomb of the False Door, not far from the large decorated stelae. 77 Moreover, a spearhead from the above mentioned Chamber B in the Tomb of the Brick Arches has a distinctive shape characterized by the change in width of the blade forming concavities on its sides. 78 This shape is reminiscent of fourth-fifthcenturies CE spear-heads from Post-Meroitic royal tombs at Qustul-Ballana 71 Conti Rossini 1928, 111-112, 133, 141;Marrassini 2014, 44-45;Phillipson 2012, 84. 72 Bernand et al. 1991, 241-245, no. 185, ll. 2-3, 246-250, no. 185 bis, ll. 3-4, 250-254, no. 186, ll. 4-5, 255-258, no. 187, ll. 3-4, 258-261, no. 188, l. 5, 363-367, no. 270, ll. 5-6, 367-370, no. 207 bis, ll. 4-5; see also Phillipson 2012, 84, 95-96. 73 Munro-Hay and Juel-Jensen 1995, 38. 74 Munro-Hay 1995Munro-Hay and Juel-Jensen 1995, 160-161, Type 67; see also Phillipson 2012, 87, fig. 25. 75 Munro-Hay 1989Phillipson 2000, I, 111, figs. 92 j-q, fig. 93; see also Manzo 1998, 46-47. 76 Phillipson 2012, 87. 77 Chittick 1974Munro-Hay 1989, 319, fig. 18.9. 78 Phillipson 2000, I, 111, fig. 93. and at El-Hobagi. 79 Indeed, as in the case of many of the spear-heads from Qustul-Ballana and El-Hobagi, the one from the Tomb of the Brick Arches is characterized by a rib in the middle and is more than 60 cm in length, which would have made it very fragile if used in battle. Therefore, they may be weapons used as rank markers and display or for ceremonial purposes, as was the case in the Meroitic and Post-Meroitic funerary contexts. 80 Thus, apparently around the fourth century CE not only were the military connotations considered crucial by both the Aksumite and Nubian-Sudanese elites, 81 but also some artefacts used to display and communicate this were shared by the two contexts. It is possible that this indicates contact between them, 82 which may have arisen from the complex relationships between Aksum and the peoples of the Nile valley, most likely only partially depicted in the royal Aksumite inscriptions, 83 and otherwise shown by very few and much debated archaeological elements. 84

Final Remarks
To sum up, the large decorated stelae were a pivot of the urban space in Aksum and may have been the location for ceremonies aimed at transmitting the ideas and concepts to the people, the Aksumite rulers regarded as crucial. In this respect, the stelae may have had the ultimate goal of forming Aksumite identity. As far as can be seen, the symbols used to communicate these ideas were elements already rooted in both local traditions (the idea of funerary stela itself), and selected foreign elements (the vine-leaf pattern on the 'Altarplatte' of Stela 3, the plaque with the head of Gorgon from the Tomb of the Brick Arches, and perhaps some kinds of weapons represented on the stelae and being found among the grave goods). In a similar way, during the same phase, the Greek language and alphabet, and South Arabian script were used beside the Gǝʿǝz script and language in the so-called trilingual royal inscriptions, to express typical Aksumite concepts. 85 The challenge for scholars now is to understand why the Aksumite rulers selected these symbols instead of all the others available for the purpose of express-79 Lenoble 2018, 119-120, fig. 64, 193, fig. 65, 198, 213-219, fig. 109, no. 55, fig. 110, no. 58-59;Manzo 1998, 46-47; see also Phillipson 2000, I, 111. 80 Francigny 2016, 126, 128;Lenoble 2018, 225. 81 For the Nubian context see lastly Lenoble 2018, 241-247. 82 See also Manzo 1998, 51. 83 Hatke 2013Marrassini 2014, 230-246. 84 See lastly Phillips 2014. 85 Conti Rossini 1928Marrassini 2014, 29, 36-37. ing the ideas and concepts to their people, and ultimately construct an Aksumite shared identity.

Appendix A: Again on the Snakes in Ancient Ethiopia
In a recent article published in this journal, I suggested, based on its representation on some archaeological finds and monuments, that the snake may have had an ideological relevance in Proto-Aksumite and Aksumite contexts, and these finds go some way towards confirming the well-known legend of a giant king snake ruling over Ethiopia. 86 Also in the light of what was said in the present article about the Gorgon head from the Tomb of the Brick Arches in the May Ḥǝǧǧa stelae field, it may be interesting here to highlight several large pieces of possible snake figurines collected from the pre-Aksumite Area D at the site of Kidanä Mǝḥrät, a domestic area investigated by the British expedition at Aksum. 87 This find may support the hypothesis that snakes already possessed an ideological meaning in pre-Aksumite times, and may thus represent the continuity of ideological values from the pre-Aksumite into the Proto-Aksumite and Aksumite phases, in addition to the use of the stelae as funerary marker already stressed in the article, and to other aspects in the monumental architecture and material culture which have already been noted. 88 Aethiopica 23 (2020) 20 Fig. 1 View of the May Ḥǝǧǧa stelae field at Aksum from east (Archives of the Archaeological Expedition at Aksum of the Università degli Studi di Napoli "L'Orientale"). The photo was taken in October 2006 (before the re-erection of stela no. 2 according to the DAE numeration) from the top of the May Qoḥo hill, the Betä Giyorgis hill is the one dominating the stelae field from north. Fig. 2 The 'Altarplatte' of the stela of Ǝnda Iyäsus, no. 4 in the DAE numeration, note the offering basins carved on its upper surface (Archives of the Archaeological Expedition at Aksum of the Università degli Studi di Napoli "L'Orientale"). They are in the form of two-handled vessels.

The Great Aksumite Decorated Stelae
Aethiopica 23 (2020) 21 Fig. 3 Detail of the upper surface of 'Altarplatte' of the big standing stela, no. 3 in the DAE numeration (Archives of the Archaeological Expedition at Aksum of the Università degli Studi di Napoli "L'Orientale"). Note the vine-leaf with frieze of bunches of grapes decorating its edges and the carved offering basins whose reuse as mortars has completely effaced their original shape, some carved gäbäṭa board are also visible on the slab.
Aethiopica 23 (2020) 22  . Note the wall delimiting the platform on top of which the stelae were erected facing a large empty space. Fig. 6 The top of the large standing decorated stela, no. 3 in the DAE numeration, with the representation of the 'monkey's heads' and several types of windows arranged in superimposed rows (from Phillipson 1997). Note the iron attachments for a metal plaque in a recessed sector and the concavities on the sides demarcating the top of the stela. Fig. 7 The top of a big decorated stela, no. 4 in the DAE numeration, with the carved representation of the 'monkey's heads' and of windows, the recessed sectors for metal plaques and four concavities on the sides demarcating the top of the stela (from Chittick 1974). Note the carved spears on one side and a rounded element, most likely a shield, on the other.

Summary
The article presents a review of the architectural and iconographic features of the big decorated Aksumite stelae in the May Ḥǝǧǧa stelae field at Aksum. Their location in the urban setting of ancient Aksum is scrutinized alongside their ceremonial and ideological function. The origin and meaning of the different features and decorative patterns characterizing the stelae are focused upon. It is suggested that these attributes may reflect ideological traits regarded as crucial by the Askumite kings and the elite in the first centuries CE. Moreover, it is demonstrated how some of these features are rooted in the local traditions, while others are related to the intense interactions the ancient Aksum had with neighbouring regions, such as the Mediterranean area and South Arabia. In particular, a new interpretation is proposed for the very distinctive outline of the top of the monuments: it may have been shaped after a specific type of shield also occurring in Meroitic and Post-Meroitic Nubia. Finally, it is suggested that the role these monuments may have played in ceremonies aimed at shaping the Aksumite identity.